Seminar: The Design of Mechanical Intelligence

Abstract

The concept of mechanical intelligence refers to the ability of a mechanism to respond/react to the environment or automatically perform some function(s) without guidance from a controller. This approach is beneficial for a variety of quasi-static and dynamic applications including human movement enhancement (exoskeletons, orthoses, prosthetics), manufacturing equipment (high-speed machinery, grippers), and robot locomotion (legs, wings, suspensions).  For example, legged robots propel themselves through short ground contact periods where the forces that may be exerted are subject to the limits of the motor/controller setup. Instead, if passive mechanics were to automatically drive the desired forces/motion, the online effort to control low level mechanics is transplanted to the offline effort to design mechanical intelligence. But due to the complexity of underlying problems, creating useful design tools is a challenge. Nonetheless, this approach is capable of removing the power and bandwidth limitations associated with traditional serially actuated mechanisms.  In this presentation, the design of mechanical intelligence is demonstrated to realize the next level of dynamic machines.

About the speaker

Mark Plecnik is a postdoctoral scholar at the University of California, Berkeley, in the Biomimetic Millisystems Lab. He received his PhD in mechanical engineering from the University of California, Irvine in 2015. His interests are in computational mechanical design, particularly for robot locomotion or other dynamic systems. Consequently, his research also lies in studying homotopy continuation as the main mathematical toolset for design algorithms. Plecnik’s work won a best paper award at the 2015 ASME International Design Engineering Technical Conferences and has appeared on the cover of Science Robotics.

Hosted by Prof. Haijun Su.